Voting on the Amendment to Restrict NSA Data Collection

(Revised, July 30, 2013)

In an America where partisan polarization has reached levels unseen since Reconstruction, the vote on July 24th to restrict funding for some of the surveillance activities conducted by the National Security Agency offered an welcome counterpoint.

Representative Justin Amash (R-MI) proposed an amendment to the defense appropriation bill that would have blocked funding for the wholesale collection of telephone calling records, so-called “metadata,” by the NSA. Despite heavy lobbying from the President, senior members of his Administration, and the leadership in Congress who all favored the amendment’s defeat, in the end it lost by only twelve votes, 205-217.


Fully a majority of Democrats voted in favor of ending the policy of indiscriminate NSA collection of telephone records despite admonitions by the President that the program be maintained while undertaking “a reasoned review of what tools can best secure the nation.”  Over in the House both Speaker John Boehner and former Speaker Nancy Pelosi voted against the amendment, with Boehner delivering a especially forceful defense of the current surveillance programs.  (Usually the Speaker refrains from voting in roll calls, so Boehner must have thought this an important vote to cast.)  Despite their leader’s strong opposition to the amendment, about two out of five Republican Members also voted to end the surveillance program.

All sorts of theories have been suggested by the pundits for the pattern of votes on this amendment.  Since the usual left-right dimension fails to predict the divisions within the parties, we must look elsewhere for explanations.  One common theory is that surveillance and other civil liberties issues form a “second dimension” of political conflict.  This view pits civil libertarians on both the left and the right against those who believe that, in some circumstances, the demands of national security outweigh individual privacy rights.

To examine these theories we would need a way to determine the position of each Member of Congress on these dimensions of political conflict.  Luckily some smart political scientists have devoted quite a bit of their professional lives to developing a method of scoring legislators based on the roll call votes they cast.  The scores for every Congress through the one that just ended are available on the Voteview site.  That limits my analysis to the 342 Members who served in the preceding Congress and voted on the Amash amendment.*

In an earlier blog posting, these analysts suggested that the second dimension their method identifies might be correlated with voting on domestic surveillance issues.  In a reply to a posting by Nate Silver, they analyzed two votes from the last Congress, one to renew the Patriot Act in 2011, and one to renew FISA in 2012.  In this posting they suggested that the mysterious “second dimension” might help differentiates Members’ votes on security and surveillance issues:

Previous conflicts like civil rights in the mid-twentieth century and bimetallism in the late-nineteenth century emerged as “second dimension” conflicts, meaning that the underlying (first) liberal-conservative dimension is insufficient to explain cleavages over these issues. It is too early too tell whether the second dimension is truly capturing an “establishment vs. outsider” divide over issues like domestic surveillance and the debt ceiling in contemporary congressional voting or merely fitting noise on a special subset of roll call votes, but the evidence so far is suggestive that the second dimension may be real and important.

However Voteview’s analysis this week of the Amash vote found little evidence of this second dimension influencing Members’ votes on the issue.  Rather they saw a pattern of support among more extreme Members on both sides of the aisle.  Using the approach I describe in detail in the Technical Appendix, we can plot the estimated probability of voting for the Amash amendment by ideological position like this:


This chart is based on estimating the influence of ideology separately for Democrats and Republicans.  The median members of each party are represented by blue and red vertical lines respectively.  The model predicts the chances that the median Democrat would vote for the amendment at about two-to-one; the median Republican opposes the amendment with about the same two-to-one odds.  The much narrower range of opinion within the Democratic delegation helps explain the much steeper slope we see for Democrats.  The horizontal lines along the X-axis represent the range of ideology scores within one standard deviation of the median. A change of 0.1 on the ideology measure represents a larger difference of opinion within the Democratic delegation than it does among the Republicans.

We can also identify two groups, one within each caucus, that display higher predicted rates of voting for the amendment.  Among Democrats, those Members elected alongside President Obama were substantially more likely to favor restricting the NSA.  Of the fourteen Members elected for the first time in 2008, twelve of them voted in favor of the amendment including the most centrist member of the group, Bill Owens of upstate New York.

Across the aisle, the equivalent contingent of 2008 Republicans were also more likely to favor the amendment as were those Republicans elected for the first time in the 2010 landslide.  There are many more Members in this group, 88 in all, with DW-NOMINATE scores running from 0.386 for Jon Runyan of New Jersey to a 1.0 for South Carolina’s Mick Mulvaney, both members of the class of 2010.  (Mulvaney voted in favor of the amendment as did the only other Member of the House scored to Mulvaney’s right, Jim Sensenbrenner of Wisconsin, the author of the PATRIOT Act.  The New York Times reports that Sensenbrenner’s brief floor speech in favor of the Amash amendment strongly influenced wavering Members.)

This analysis shows clearly the generational divide within both parties, and particularly within the Republicans.  The few Democrats elected in 2010 showed no greater support for the amendment than other Democrats with equivalent ideological beliefs.  For Republicans, those elected in 2008 and those elected in 2010 were both more likely to favor NSA restrictions at a statistically identical rate.

Two other factors not shown in the graph also influenced votes on the Amash amendment.  Members of both the House Armed Services Committee and, especially, the Select Intelligence Committee were substantially more likely to oppose the Amash amendment than Members with equivalent ideological positions.

For details of the model, see the Technical Appendix for this posting.


* People who know how high House re-election rates are might be surprised that only 349 Members of those voting or absent Wednesday night were returning incumbents.  The turnover rate for the 113th Congress was the highest it has been for decades.  The rate of retirements skyrocketed, while the re-election rate for incumbents declined. Redistricting played a role especially in states like California where incumbents were forced to run against each other. The substantial influx of newcomers in 2013 led one observer to call this Congress one of the “most inexperienced in history.”